[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16083231#comment-16083231
]
Rajesh Balamohan commented on TEZ-3754:
---
This is not an infinite loop similar to the one which got
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16083137#comment-16083137
]
Prasanth Jayachandran commented on TEZ-3754:
Can we have a counter for shuffle
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16083027#comment-16083027
]
Siddharth Seth commented on TEZ-3754:
-
[~rajesh.balamohan] - what is the corner case in which this is
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16081281#comment-16081281
]
Rajesh Balamohan commented on TEZ-3754:
---
Agreed that this would block the cases which could have
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16080529#comment-16080529
]
Jonathan Eagles commented on TEZ-3754:
--
The removal of drop behind also helps in the broadcast fetch
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16080500#comment-16080500
]
Jason Lowe commented on TEZ-3754:
-
Note that we mitigated some of the effects of shuffle hammering in
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16080464#comment-16080464
]
Jason Lowe commented on TEZ-3754:
-
bq. do you why mapreduce explicitly turned this off?
As Zhiyuan
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16080445#comment-16080445
]
Jonathan Eagles commented on TEZ-3754:
--
[~jlowe], do you why mapreduce explicitly turned this off? I
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16079940#comment-16079940
]
Rajesh Balamohan commented on TEZ-3754:
---
Thanks for the review [~aplusplus]. Pinging the host in
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16079913#comment-16079913
]
Zhiyuan Yang commented on TEZ-3754:
---
One side-effect is single large fetch can block other small fetches,
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16068038#comment-16068038
]
Rajesh Balamohan commented on TEZ-3754:
---
[~sseth], [~jeagles], [~aplusplus] - Can you plz review when
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16041503#comment-16041503
]
Gopal V commented on TEZ-3754:
--
[~jeagles]: Netty4 uses the kernel defaults instead, so in LLAP we're reading
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16041022#comment-16041022
]
Jonathan Eagles commented on TEZ-3754:
--
Agree about the intention of the jira. However, can you comment
[
https://issues.apache.org/jira/browse/TEZ-3754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16040766#comment-16040766
]
TezQA commented on TEZ-3754:
{color:red}-1 overall{color}. Here are the results of testing the latest
14 matches
Mail list logo